microsoft/mcp-gateway

MCP Gateway is a reverse proxy and management layer for MCP servers, enabling scalable, session-aware stateful routing and lifecycle management of MCP servers in Kubernetes environments.

What it solves

MCP Gateway provides a scalable, stateless reverse proxy and management layer for Model Context Protocol (MCP) servers. It solves the problem of managing the lifecycle (deployment, updates, and deletion) and routing of MCP servers within Kubernetes environments, while adding enterprise-grade authorization and observability.

How it works

The project operates through two primary planes:

  • Control Plane: A set of RESTful APIs used to deploy and register MCP servers (as "adapters") and specific tools. It manages the backend Kubernetes cluster, handling the deployment of server images and storing metadata in Redis or Cosmos DB.
  • Data Plane: A reverse proxy that routes incoming MCP requests. It supports direct routing to specific adapters via /adapters/{name}/mcp or dynamic routing via a Tool Gateway Router. The router acts as an intelligent MCP server that analyzes tool calls and forwards them to the correct registered tool server.

Who it’s for

Developers and platform engineers building AI applications that rely on the Model Context Protocol and need to deploy these tools at scale in Kubernetes, requiring robust access control (Azure Entra ID) and centralized management.

Highlights

  • Stateless Routing: Independently routes each request without requiring transport-session affinity.
  • Dynamic Tool Routing: A specialized router that directs tool execution requests to the appropriate server based on tool definitions.
  • Kubernetes-Native: Uses StatefulSets and headless services for deployment.
  • Integrated Management Portal: Includes a React SPA for managing adapters and tools and testing them via a JSON-RPC console.
  • Agent Support (Preview): Optional ability to run LLM-driven agents that can call registered MCP tools or other agents.

Related

  • Project
    agentic-community/mcp-gateway-registry

    A governed control plane and gateway for AI assets, providing centralized discovery, access control, and audit trails for MCP servers, agents, and model inference endpoints.

  • Project
    containers/kubernetes-mcp-server

    A native Go implementation of the Model Context Protocol (MCP) server that allows AI agents to directly manage and troubleshoot Kubernetes and OpenShift clusters.

  • Project
    microsoft/azure-devops-mcp

    An MCP server that gives AI agents direct access to Azure DevOps data and tools, enabling them to manage work items, repositories, and pipelines.

  • Project
    apify/mcpc

    mcpc is a lightweight, cross‑platform CLI that gives shell access to any Model Context Protocol (MCP) server, supporting full MCP features, persistent sessions, OAuth 2.1 security, JSON scripting, dynamic tool discovery, and experimental x402 payments.

  • Project
    supercorp-ai/supergateway

    A transport gateway for Model Context Protocol (MCP) servers that enables stdio-based servers to be run over SSE, WebSockets, or Streamable HTTP for remote access and aggregation.